Knowledge Tracing: A Survey

Abstract

Humans’ ability to transfer knowledge through teaching is one of the essential aspects for human intelligence. A human teacher can track the knowledge of students to customize the teaching on students’ needs. With the rise of online education platforms, there is a similar need for machines to track the knowledge of students and tailor their learning experience. This is known as the Knowledge Tracing (KT) problem in the literature. Effectively solving the KT problem would unlock the potential of computer-aided education applications such as intelligent tutoring systems, curriculum learning, and learning materials’ recommendation. Moreover, from a more general viewpoint, a student may represent any kind of intelligent agents including both human and artificial agents. Thus, the potential of KT can be extended to any machine teaching application scenarios which seek for customizing the learning experience for a student agent (i.e., a machine learning model). In this paper, we provide a comprehensive survey for the KT literature. We cover a broad range of methods starting from the early attempts to the recent state-of-the-art methods using deep learning, while highlighting the theoretical aspects of models and the characteristics of benchmark datasets. Besides these, we shed light on key modelling differences between closely related methods and summarize them in an easy-to-understand format. Finally, we discuss current research gaps in the KT literature and possible future research and application directions.

Knowledge tracing: A brief overview of the advances done since 2021 using deep learning techniques

“…This year, (Abdelrahman et al, 2022), published another survey covering a broad range of methods starting from the early attempts to the recent stateof-the-art methods, they highlighted the theoretical aspects of models and the characteristics of benchmark datasets, herein neither a review of datasets, assessment methods not methods that are not based on deep learning are contemplated, however it is contrasted the taxonomy proposed for deep knowledge tracing with the one proposed by (Q. Liu, Shen, et al, 2021) we shed light on key taxonomizing differences between closely related methods and summarize them.…”

“…1, which contains two taxonomies; (a) (Q. Liu, Shen, et al, 2021), and (b) (Abdelrahman et al, 2022). To trace the timeline of the development is introduced a representative publication of each category, thus firstly notice that both, taxonomies (a) and (b), are very similar, lets analyse the categories on more detail:…”

“…1: A taxonomy of deep learning knowledge tracing models (a)(Q. Liu, Shen, et al, 2021) , (b) (Abdelrahman et al, 2022). All these previously mentioned models are solutions based in deep learning technics however that do not mean that models based on machine learning have also been constantly evolving and producing additional solutions, however those solutions land out of this work scope, for those who are interested on them is highly recommended to check them out.…”

Prerequisite Structure Discovery in Intelligent Tutoring Systems

“…Therefore, a relevant evaluation of KS discovery algorithms must use data with exercise sequences agnostic of the domain knowledge, to remove the possibility that the algorithm discovers the KS owing to this expert knowledge instilled in the exercise sequences, and to ensure that KS discovery methods capture information from user performances even in the absence of prior domain knowledge. Finally, evaluation of the discovered KS requires comparing with the ground truth KS -which is absent from publicly available datasets [10].…”

A study of adults' orientations to learning at the national open university of Taiwan

“…This paper proposes an Adaptive Learning Path Navigation (ALPN) system that recommends learning materials to students according to their knowledge states, i.e., the mastery level of concepts in a subject. The system employs a Knowledge Tracing (KT) model to quantify students' current knowledge states [26], followed by a decision-making model that recommends tailored learning materials. As students complete the learning materials, the KT model updates their knowledge states by assessing their responses.…”